Evolutionary and functional relationships of insect immune proteins Marco Fabbri Supervisors: Prof. Otto Schmidt Dr. lJlrich Theopold A thesis submitted for the degree of Doctor of Philosophy in the Department of Applied and Molecular Ecology The University of Adelaide May 2003
91
Embed
Evolutionary and functional relationships · 2015. 5. 25. · Evolutionary and functional relationships of insect immune proteins Marco Fabbri Supervisors: Prof. Otto Schmidt Dr.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Evolutionary and functional relationships of
insect immune proteins
Marco Fabbri
Supervisors:
Prof. Otto Schmidt
Dr. lJlrich Theopold
A thesis submitted for the degree of
Doctor of Philosophy
in the Department of Applied and Molecular Ecology
The University of Adelaide
May 2003
Contents
Innate immunity..........
Recognition molecules
Cellular response
Hemolymph coagulation.............
Exocytosis of hemocytes mediated by lipopolysaccharide
Coagulation mediated by lipopolysaccharide.
Coagulation mediated by beta-glucan .........,.
Lectins
Glycosylation................
Glycosylation in Insects
Glycosyltransferases .....
Mucins
Hemomucin ..................
Strictosidine synthase
Drosophila melanogaster as a model
4
10
1.3
15
.15
6
17
t9
2T
22
24
26
26
I
t7
A Lectin multigene family in Drosophila melanogaster
Introduction
Materials and Methods
Sequence similarity searches
28
28
..29
29
Results 30
Novel lectin-like sequences in Drosophila 30
Discussion .....31
34Figures
An immune function for a glue-like Drosophila salivary protein 39
Introduction 39
Materials and Methods.............. .....41
Flies .. 4r
Hemocyte staining with lectin..
Electrophoretic techniques .......
N-terminal sequencing of pl50
Pl50 - E. coli binding
RNA extraction.
In situ hybndizations ........,
Isolation of Ephestia ESTs
RT PCR
Relative quantitative PCR
41
4I
42
42
43
43
43
44
44
Results
II
.46
p150 is 171-7 47
49Discussion......
Figures.... 53
Animal and plant members of a gene family with similarity to alkaloid-
Introduction 58
S equence smilarity searches
Insect cultures
Preparation of antisera
Immunoblotting and Immunodetection of Proteins........
Radiolabelling and Purification of DNA Probes
59
59
59
60
60
6t
61
Hybridization conditions
Northern blots ................
Results 62
Novel strictosidine synthase and hemomucin .......... -.'.'.-.......-'.-'...62
Discussion .....64
Figures 67
III
Summary
Innate immunity has many features, involving a diverse range of pathways of immune
activation and a multitude of effectors-functions. This thesis project examined different
aspects of innate immunity. In the first part a novel gene family of putative C-type lectins
is presented, which may have developmental and immune functions. In the second part,I
explored the possible ancestral origin of immune-effector proteins by investigating
salivary gland and silk proteins involved in coagulation. In the last part, novel immune
genes with similarity to strictosidine synthase are presented and their role as antifeedent
is discussed.
Glycodeterminants play an important role in mediating cellular and cell-substrate
interactions during development and immune-related reactions enabling an organism to
distinguish self determinants from non-self or modified-self determinants. The most
studied sugar recognition molecules are lectins. They have a wide range of binding
activities and they are organized in multigene families. Here I describe a group of D.
melanogaster geîesthat arc possible members of the C-type lectin family.
Characterization of a novel D. melanogaster hemocyte mucin revealed a gene-locus I71-
7, which was identified as a salivary (labial) gland protein. These data suggestthatlTl-7
is expressed similarly to hemomucin and may take part in hemol¡rmph coagulation and
entrapment of microorganisms. To test whether labial gland proteins are expressed in the
immune system of other insects, I studied two lepidopteran silk proteins and found them
1
to be expressed by immune tissues as well. The implications of labial gland secretory
protein involvement in coagulation and its role in insect immunity are discussed.
I used conserved protein domains of the Drosophila immune receptor hemomucin to
identify novel members of a gene family which have similarity to strictosidine synthase
(SS), one of the key enz¡rmes in the production of monoterpene indole alkaloids. In
addition to the first animal member of the family described previously (hemomucin) a
second D. melanogaster member could be identified, which appears to differ in
subcellular distribution from hemomucin. In Arabidopsis thaliana, SS-like genes form a
multigene family, compatible with a possible function as antifeedants and antibacterial
compounds. In Caenorhabditis elegans, two members could be identified and one
member each in Mus musculus and Homo sapiens.Interestingly, the human SS-like gene
is strongly expressed in the brain, the very organ many of the indole alkaloids act upon.
2
Introduction
Innate immunity
Genetically 'hard-wired' defense systems termed "innate immunity" are present in all
individuals at birth and include all responses that are independent of specialized
immunocompetent T and B Lymphocytes. Innate immunity provides a first line of
defense against pathogens in vertebrates and is the only defense system in invertebrates.
In vertebrates it limits the infection and provides signals required for the development of
the adaptive immune response (Biron et a1.,1989; Fearon and Locksley,7996).
Host defense in vertebrates is an intricate interplay between innate and adaptive
responses, involving mechanisms that reflect the diversity of pathogens that infect them.
It is importantthat these defense mechanisms be present in readiness at all times and
rapidly activated upon infection to allow for the destruction of the infectious agents. It is
also obvious that inactivation of pathogens must proceed without harming the host itself.
Following this idea, Charles Janeway argued in 1989 that innate immunity must operate
by means of receptors that have been selected over evolutionary time to recognize highly
conserved and widely distributed features of pathogens, especially features that are not
found on the cells of host organisms. He called these features "microbial pattems" and
coined the term of "pattem recognition receptors" (PRRs) for the receptors bbatrecognize
them (Janeway, 1989).
PRRs are carried by particular types of cells, such as macrophages, natural killer (NK)
cells, and probatily also epithelial and endothelial cells in the lung, kidney, skin, and
gastrointestinal tract (Unanue, 7984; 'Wright,
1991). They generally have a broad ligand
specificity and fail to discriminate between pathogen species. As opposed to clonal
molecules such as the immunoglobulins (Igs), their expression is not induced in response
4
to specific pathogen groups, but rather to broad classes of microbes such as Gram-
negative or Gram-positive bacteria, and/or fungi. The innate immune response does not
generate immunological memory, nor does it trigger long-lasting protection against
disease. However, absence or genetic defects in these defense mechanisms can lead to
recurrent infections, demonstrating their importance (Super et aL.,1992).
The recognition of pathogens by PRRs triggers their engulfment by immune competent
cells leading ultimately to the destruction of the pathogens (Unanue, 1984). PRRs also
participate in the elimination of tissue debris that accumulates during infection,
inflammation, and wound repair (Akbar et al., 1994; Savill, 1997a). Importantly, this
engulfinent process, called "phagocytosis", is also important during development, as it
pafücipates in the clearance of cells that undergo programmed cell death (PCD, or
apoptosis) (V/yllie et al., 1980; Raff, 1992; Savill, 1997b). Rapid phagocytosis of the
apoptotic corpse by specific receptors prevents the release of secondary immune signals,
which could affect the homeostasis of neighboring tissues (Wyllie et al., 1980; Raff,
1992). Recent studies have also highlighted the importance of phagocytosis receptors
with specificity for apoptotic cells in the suppression of inflammation (Akbar et a|.,1994;
Savill, 1997a). Furthermore, recent data suggest that failure to dispose of apoptotic
corpses could ultimately result in the activation of immune responses against self-
antigens, which may contribute to autoimmune diseases such as lupus (Laderach et al.,
1998; Botto et aL.,1998).
Insects possess a complex and efficient system of biological defense against pathogens
and parasites. This system comprises three different means: the integument and gut as
first line of defense to infection, the responses within the hemocoel when these barriers
are breached and the induced synthesis of antimicrobial peptides and proteins from
various tissues, such as the fat body.
5
Recognition molecules
In order to identify a potentially damaging object or organism an immune system must be
capable of recognising diagnostic features at a molecular level. Furthermore, this
recognition process must be sufficiently precise in order to allow a defence reaction to
take place in an accurate, effective and controlled fashion. Therefore, the process of
recognising potentially damaging structures must be considered as the key event for a
specific immune response.
Since the insect immune system is not adaptive, an immune response relies on a fixed
number of recognition molecules that are specific for common microbial epitopes
(Hultmark, 1993). They are thought to be representative for a broad spectrum of
microbes, as indicated by the term "pattem recognition" (Janeway,1994).
However, pattern recognition is probably not a complete description of innate immune
recognition processes since abiotic and chemically inert particles (e.g. those made from
Nylon), are able to stimulate immune responses in vivo (Salt, 1965). Therefore, it is
conceivable that physical properties, like the net charge and the hydrophobicity of a
surface, play a crucial role in immune reactions, whether or not a foreign object is
recognised as "non-self' (Lackie, 1988). In fact, objects with neutral surfaces, when
injected into the hemocoel of cockroaches, provoked a less intensive response than
charged surfaces (Lavine and Strand, 2001). On the other hand, negatively charged
surfaces were not encapsulated by hemocytes of certain caterpillars or locusts (Lackie,
1986). A few potential recognition molecules have been identified to date and some of
them are described in the following sections.
6
Hemolin, or P4, a 48kDa protein found in the hemolyrnph of Hyalophora cecropia and
Manduca sexta is considered as one of the candidate recognition molecules (Sun et al.,
1990; Ladendorff and Kanost, 7991). Interestingly, this molecule contains four intemal
repeats showing homology to C2-type immunoglobulin-like domains, indicating that it is
a member of the immunoglobulin superfamily. Within this superfamily, hemolin shows
the closest similarity to cell adhesion molecules, particularly to the insect proteins
neuroglian and amalgam and the vertebrate proteins NCAM and Ll (Sun er al.,1990).In
unchallenged larvae, it is present at low levels - a prerequisite for putative recognition
molecules. After bacterial infection, however, the concentration of hemolin increases up
to l8-fold (Andersson and Steiner, 1987). Sun e/ al. (1990) could show that hemolin
binds to the surface of bacteria, forming a complex with two other proteins. However,
hemolin itself seems to display no direct antibacterial activity supporting the idea that
hemolin functions as a recognition molecule. In fact, several lines of evidence indicate
that hemolin is involved in the regulation of hemocyte behaviour. Upon immune
challenge in vitro, hemolin binds to hemocytes and has the ability to inhibit the
aggregation of hemocyt es (Zhao and Kanost , 7996; Ladendorff and Kanost , l99l; Kanost
et al., 1994; Lanz-Mendoza et al., 1996).In addition to its role in regulating hemocyte
adhesiveness, hemolin is capable ín in vitro experiments to stimulate the phagocytic
activity of insect blood cells (Lanz-Mendoza et al., 1996). However, the in vivo function
of hemolin remains to be determined. Interestingly, (Bettencourt et al., 1997) found
evidence of a 52kDa membrane form of hemolin associated with hemocytes, indicating a
multifunctional role of this putative recognition molecule in humoral and cellular
immunity.
7
For another class of recognition factors, the lectins, it has been clearly shown that they
play a role in insect immunity. However, their precise mode of action has still to be
resolved. In general, lectins are widely occurring proteins or glycoproteins, sometimes
with multiple binding sites, which specifically recognize certain carbohydrate moieties.
They are inducible and show no enzyrnatic activity. Due to their ability to agglutinate
vertebrate erythrocytes, bacteria, and other microorganisms, they are also referred to as
agglutinins (Götz and Boman, 1985). In insects, they are found to be associated with the
plasma membrane of hemocytes and as soluble hemol¡rmph proteins (Ratcliffe, 1993b).
For various lectins it has been shown that lectin-mediated recognition or clumping of
foreign objects makes these objects susceptible to both cellular and humoral defense
reactions. For example, a lectin from the grasshopper Melanoplus dffirentialzs, described
by (V/heeler et al., 1993) was shown to display an opsonic activity against fungal
blastospores.In Periplaneta americana, two lectins, named Periplanetq lectin (Kawasaki
et al., 1993) and LPS-binding protein (Jomori and Natori, 1992), respectively, have been
isolated that participate in clearing of bacteria from the hemol¡.mph. Both of them
recognise lipopolysaccharide (LPS) from the outer membrane of gram-negative bacteria
and are related in sequence to the C-type lectins of vertebrates. Another, galactose-
specific lectin from the hemolymph of the flesh fly Sarcophaga peregrina, was found to
play an active role in the lysis of sheep red blood cells injected into the hemocoel
(Komano and Natori, 1985). These few examples from invertebrates show that lectins
display opsonic activity and can be considered as recognition molecules capable of
eliciting immune responses. Likewise C-type lectins in mammals are also involved in
pattern recognition of microorganisms (Hoffmann et al.,1999).
8
Another important group of recognition molecules are proteins involved in the activation
of the phenoloxidase pathway. The end product of this pathway, melanin, plays an
important role in the sclerotisation and tanning of the insect cuticle (Anderson et al.,
1935) and is responsible for sealing off cellular capsules that have been formed around an
foreign object in the hemocoel (Salt, 1970). As exemplified in the case of already
mentioned lectins that bind LPS, sugar-determinants exposed on the surface of an foreign
object constitute important signals for the identification of bacterial microbes. In addition
to LPS from gram-negative bacteria, the peptidoglycans from gram-positive bacteria and
the mannans or beta-1,3-glucans from fungi have been described to evoke immune
reactions, including the activation of the prophenoloxidase activating system (Sugumaran
and Kanost , 1993; Söderhäll et al., 1994). Ashida and colleagues were able to purify two
soluble proteins from the hemolymph of the silkworm Bombyx mori, one of which is
specific for peptidoglycans from bacterial cell walls, and the other for beta-l,3-glucans
from fungal cell walls (Ashida and Yamazaki, 1990). When bound to their respective
ligands, these proteins activate a proteolytic cascade, named prophenoloxidase activating
system, that leads to the activation of the enzpe phenoloxidase which is generally
considered to be the key enzyme for the s¡mthesis of melanin (Sugumaran and Kanost,
1993). Since many cellular capsules are melanised, the question can be asked whether PO
or prophenoloxidase þroPO), the non-activated form of PO, serves as a recognition
molecule triggering the encapsulation reaction. In this context (Rizki and Rizki, 1990)
could show that in larvae of proPO-dehcient mutants of Drosophila melanogaster
parasitoid eggs are encapsulated but not melanised. From this result they concluded that
phenoloxidase is involved in the cross-linking and melanisation of capsules, but not in the
recognition of foreign objects in Drosophila lawae.It seems that the recognition events
9
leading to the production of melanin and the ones leading to encapsulation are separate
events.
Cellular response
The cellular defence reactions are mediated by the insect blood cells in the hemolyrnph,
the hemocytes. Hemocytes from different insects show large variations regarding their
morphological and functional characteristics. Therefore, hemocyte classification has
proven to be difficult. However, two types of hemocytes, the plasmatocytes and the
granulocytes, are generally thought to be the most important ones involved in cellular
defence reactions (Ratcliffe, 1993a). Both plasmatocytes and granulocytes aÍe
polymorphic and variable in size. A distinguishing feature of granulocytes is that they
possess numerous cytoplasmic granules. Another characteristic of this hemocyte type is
the presence of a microtubule band near the periphery of the cell. In contrast,
plasmatocytes are agranular or contain granules in the cytoplasm, which are considerably
finer and less electron-dense; peripheral microtubule bands are absent (Gupta, 1991).
The invasion of foreign particles into the insect hemocoel evokes dramatic changes to the
hemocyte population, involving surface morphology and adhesive properties (Wago,
1980b; 'Wago, 1980a). Granular hemocytes, for example, form numerous filipodia on
their cell surface, important for entrapping foreign particles, and become increasingly
adhesive to each other and to other surfaces (Wago, 1980a). When granulocytes get in
contact with a foreign surface, they undergo degranulation, a process involving the
discharge of granules and other cytoplasmic contents into the surrounding environment.
The released components, which are generally thought to include phenoloxidase and
lectins, act as opsonins that attract additional hemocytes and, in general, invoke the
insect's defence reactions (Ratcliffe, 1993a). Nevertheless, the mechanisms by which
10
hemocytes recognise a foreign object are still the least understood facet of the cellular
defence reactions. However, recognition molecules associated with hemocytes and
present in cell-free hemolymph seem to be involved (Ratcliffe,1993a). Depending on the
number and size of the foreign particles in the insect hemocoel, three major defense
reactions can be distinguished: phagocytosis, encapsulation, and nodule formation.
The principle means of removing small foreign objects such as bacteria from the
hemolymph is phagocytosis, a process comprising recognition, attachment, and
internalisation of the foreign object by hemocytes (Gupta, 1991). Phagocytic activity has
been reported for both granulocytes and plasmatocfes. After a particle is recognized as
foreign and attached to a hemocyte, it becomes ingested by endocflosis. In this process,
the foreign particle is taken up by the hemocyte into either a coated vesicle, a membrane
bound phagosome or a pinosome, each of which eventually fuses with a lysosome.
Lysosomes contain en4rmes and antimicrobial agents that finally digest the foreign
object (Gupta, 1991). In the case of microbes, this process is believed to lead to the
release of cell wall components, like lipopolysaccharides or peptidoglycans, that in turn
are strong elicitors of antibacterial and antifungal defense reactions (Iketani and
Morishima, 1993).
Foreign objects like metazoan parasites too large to be phagocytised are surrounded by
hemocytes and become engulfed in a multi-layered cellular capsule. Similar to
phagocytosis, both granulocytes and plasmatocytes have been shown to participate in
encapsulation (Pech and Strand, 1996). However, in some species, only granulocytes are
capable of encapsulation, a property that appears to be dependent on the granulocyte
microtubule band (Gupta, 1991). The initial stage of encapsulation involves the
11
recognition of foreign antigens by granulocytes. After recognition, the granulocytes
degranulate, release stored components, and eventually lyse. These components adhere to
foreign surfaces and, in turr;', attract additional hemocytes (Gupta, 1991). The next step
leads to the formation of multiple cellular layers around the foreign object. The
hemocytes in contact with the foreign particle form the irurermost layer (Salt, 1970). They
show elongation, flattening, and lysis. Melanization occurs, a process which physically
separates and eliminates the foreign object from the hemol¡rmph. However, gradually
more hemocytes become attached, building up new layers. Finally, the capsule develops
an outer layer which does not invoke any further attachment of hemocytes (Salt, 1970;
Gupta, 1991). Interestingly, in Pseudoplusia includens, this outer layer is formed by a
monolayer of granulocytes (Pech and Strand, 1996).
When large numbers of foreign objects, such as microbes, have entered the insect's
Fig. 2: Sequences of Drosophila ORFs that sho\il homology 'with UDP-glucuronosyltransferases. The baculovirus ecdysone galactosetransferase is used as an
insect member of the UGT family for comparison. UGT35a and UGT35b are twomembers of a family of 5 antennal-specific Drosophila UGTs that have been describedrecently (Wang et al., 1999). They are almost identical to the ORFs above them(4C0064918 and 4C006491F) except for a few arnino acid differences which might be
due to polymorphisms. A consensus sequence, which is conserved amongst UGTs ispresent in all sequences of this alignment (Mackenzie et al., 1997).
F'
F'
F,
F'
F'
F,
F
tF'
F'
F'
F.
F'
F.
t,F'
F'
IrIltrIr[¡Er
tlIrIrIrtr[¡IrIrtl
IIIIIgIIIIIIIItrII
¡¡ÄÅ
Å
À¡¡ttA¡¡À
¡ilt¡
FFFFFFFFFFFFFFFFF
TTTTTTTTTETTTTT
rlrjtjrlI;rltrrjrjlrrl
t,:
ll]titi
t¡tf,'
Lt¡li:
l,i
h'
rlt;rjrllrllrjlrt.l
FFrlËLFF
FFF
HFFF
L
L
L
L
IIL
L
L
L
t-
L
tL
L
\IrlE\
H\
H
t¡H
l!
trl L
FtIt,: L
l,Lfr¡Yt l-lY¡F l¡ I,r h: Lfir[r]
H
rjltjjlr
F,
F'
F'
F'
t,t,
IrIrIr
tltltIr
irirfrfrÍrr,L
t:t,i
titi
l¡I't
F'
F'
F'
F,
F.
F,
F'
t,
rjrlllrjrjtrtlrlrj
l¡
t;
tit
t,F'
F,
F'
F'
F'
t'
h: L
tiLht L
ht L
t'F.
F,
F'
F,
lll l¡
ilTr,ilüi
Table INovel genomic fragments which contain ORFs with sequence homology to C-type
4mM CaClz, saturated with PTU). Centrifugation was carried out at 7000 rpm for
10 min at room temperature. The plasma was kept on ice, whereas the hemocyte
pellet was lysed for 10 minutes in insect Ringer containing (05% IGEPAL, 0.5%
aprotinin, and Io/o beta-amino-n-caproic acid). Centrifugation was done at 14000
42
rpm for 20 min at 4"C. The hemocyte lysate together with the plasma were added to
pre-washed E. coli, and were incubated end over end for an hour. After
centrifugation at 14.000 rpm (4'C) for 10 min, the supematant was kept and the
bacterial pellet washed 6 times with Ringer to remove unbound proteins. After
solubilization in loading buffer, samples were separated on 6% SDS-PAGE,
blotted, and analyzed using peroxidase-conjugated PNA (l¡lglml).
RNA extraction
Total RNA was extracted using an RNeasy RNA kit from Qiagen following the
supplier's instructions. DNase treatment of the RNA was performed according to
the Quiagen protocol.
In situ hybridizations
In situ hybridizations were performed as described (Theopold et aL.,1995).
Isolation of Ephestiø ESTs
ESTs with similarity to Venturia canescens virus-like particles (VLPs) were
isolated from a lambda gtll Ephestia kuehniella llbrary using a VLP specific
antiserum as described (Theopold et al., 1994). After subcloning, the clones were
sequenced and analyzed by comparison to GenBank sequences using BlastX. The
EST sequences were deposited into the dbEST database of GenBank with the
accession numbers BG69 5827-695 83 8.
43
RT.PCR
For designing primers specific for I7l-7, the annotation method was used (Wright
et al., 1996). There are some discrepancies between the genomic sequence
published by the BDGP þredicted gene CG7604) on one hand and (Wright et al.,
1996) on the other hand, most likely due to the repetitive structure of the gene.
Nevertheless, both sequences contain the protein sequence were obtained by
sequencing p150. The annotation of (Wright et al., 1996) was followed because
their sequencing was performed using exonuclease-generated clones, which
fac ilit ate s lo c alization o f rep etitive s e quenc e s.
For RT-PCR, 1 pg total RNA was reverse transcribed with 1 pg oligo (dT) 15
primer in a 20 pl reaction containing lmM of each dNTP , 40 U of RNase inhibitor
(Promega), and20 U of reverse transcriptase (AMV reverse transcriptase, Promega)
in the reaction buffer supplied by the manufacturer. RNA was denatured at 65'C for
5 min and chilled on ice before adding the other reaction components. The reaction
was carried out at 42"C for 60 min, then heat inactivated at 95"C for 5 min The
reaction was diluted to 50 ¡rl and stored at -80"C until needed. Control reactions,
without the addition of reverse transcriptase, were included in every experiment in
order to discount the possibility of DNA contamination.
Relative quantitative PCR
PCR was used to determine the accumulation of mRNA transcripts. Primers for
ubiquitin and RP49 were used to ensure that different PCRs contained equal
amounts of cDNA template. All PCRs were carried out in a PC-960 microplate
44
Thermal Sequencer System (Corbette Research). Primers used for PCR
amplification were as follows:
E p h e s t i a fibroin : GTTGATTCC GTTTAGTC CAACC and
TTCGTCATCTTCAGCAGCTTC
Galleria fibroin: GATCTTGTGCTGTGCTCTGCA and
GCACT fuA'A.GTCC CGTTCTCAA
Galleria calreticulin: CTTGCACTTAGAACGTTCCTACG and
CTCATGAAGATGCTGACTGTTCC
Ephestia ubiquitin: GCCGTTTGAATATTGAAGTCG and
TCAACAACCTGGGCAACTC
Drosophila RP49: CTTCATCCGCQCCAGTC and TCTCCTTGCGCTTCTTGG
D r o s op hi I a 17 I .7 : GAGACAC CAAGAC CAATACATC and
GAGTAGT GCTTT CAGTACAAG
Dros ophila SGS5 : CCAGGAGACGAAAATCGAAG and
GGAATGGGGAATTCAGCAAC.
Reactions were set up in a 20 p,l volume containing 2 ¡rl 10 x Taq polymerase
buffer (supplied with enzyme), 0.5 ¡rM each of two specific primers, 200 ¡rM of
each dNTP, 2mM MgCl2 and 0.5 U Taq DNA polymerase (5 U/pl; Promega).
Cycling conditions were: 5 min 94oC, followed by cycles of 94"C for 1 min/ 50"C
for 1 min I 72'C for 1 min The cycle numbers are shown in the figure legends. PCR
products were analyzed by agarose gel electrophoresis with 0.2 ¡t"glml ethidium
bromide in both gel and buffer.
After analyzing PCR products on gels, reactions were repeated with adjusted
amounts of template until the amplification products for the control (Rp49 and
45
ubiquitin) primers were equal in all samples. These adjusted amounts were
subsequently used in all other PCRs.
Results
Glycoproteins are involved in coagulation of insect hemolymph, although the
clotting process is not completely understood (Theopold et al., 2002). The
lepidopteran Galleria was one of the first model insects used to study this process
(Ratcliffe and Rowley,1979). Hemolymph bled from larvae will coagulate on glass
slides (Gregoire, 1974), where strands of extracellular material are stained by PNA
(Fig. 3 A). Following the classification of insect clot formation proposed by
Gregoire (Gregoire, 1974), Galleria clots belong to class III; showing a mixture of
fibrils and dot-like structures, although the f,rbrils are more predominant. 'When
lawae were bled into a drop of bacterial suspension, bacteria were found closely
associated with the strands and surfaces of hemocytes, and co-localized with PNA
staining (Fig. 3 A-C). This suggested that PNA-reactive proteins might be
important for microbial immobilízatiort as well as clotting. In a second lepidopteran
(Ephestia kuehniella), a pattem with prevalent dot-like structures was observed
which also bound bacteria (Fig. 3 D and E). Ephestia clots are also classified as
belonging to class III, but fibrillar components are less abundant than in Galleria.
To have more molecular genetic tools available, I switched to Drosophila as a
model system. Drosophila clots belong to Gregoire's class II, showing
predominantly membranous structures. When Drosophila lawae were bled into a
drop of bacterial suspension, the bacteria were seen associated with the PNA-
staining extracellular material (Fig. 3 F-H), broadly resembling structures observed
in Galleria and Ephestia. IJnder nonactivating conditions, the stained material is
46
seen in intracellular vesicles of different sizes (Fig. 3 I and K). Under the same
conditions, no staining was observed in preparations of non-permeabilized cells
(not shown).
In protein extracts from Drosophilahemocytes, PNA staining was concentrated in a
major protein band of l5OkDa (p150, Fig. 4 A). The same 150kDa band was
detected in a gut lysate, and a strongly stained band was also seen in salivary
glands. To investigate interactions between p150 and bactena, bacterial binding
experiments were performed using whole hemol¡rmph lysates. A substantial
fraction of p150 bound to bacteria (Fig. a B). This binding was shown to depend on
both Ca2* and sugar (Fig. a C). Inhibition was observed with galactose-containing
sugars, but not with other sugars (Fig. a C) indicating that the PNA-binding
carbohydrate determinant in p150 is necessary for binding bacteria. Thus these data
point toward the involvement of a C-type lectin-like activity with specificity for
galactose in p150 binding to bacteria. In contrast, two other mucins identified in the
hemolyrnph lysate did not bind to bacteria (Fig. a C, lower part). Since one of the
mucins (100kDa molecular mass) that did not bind bacteria is almost certainly
hemomucin, binding to bacteria is not a general property of hemolynrph mucins.
p150 is l7l-7
This suggested that p150 has unique properties by forming aggregates 'with
bacteria. This justified the molecular identification of the protein. I took advantage
of the fact that p150 is strongly expressed in salivary glands to isolate sufficient
amounts of material for protein sequencing from this tissue. The N-terminal protein
sequence obtained was unique and identifiedp150 asITI-7 (Fig. 5 A). This is one
of the proteins encoded by a locus located in the TlElate puff whose expression is
47
regulated by ecdysone (Wright et al., 1996). In contrast to other members of this
locus, which are late effector genes of the ecdysone-regulated cascade, I7l-7
belongs to the intermolt genes, which are expressed earlier during metamorphosis.
171-7 shows similarity in sequence, amino-acid composition, and in its repetitive
domain structure to Drosophila glue proteins, which are encoded by a set of
intermolt genes and are necessary to attach the pupa to a dry substrate (Wright el
al., 1996). The predicted protein sequence identified I7l-7 as a possible mucin,
with 68 potential O-glycosylation (Hansen, 1998) and 10 potential N-glycosylation
sites, numerous cysteines, and a repetitive structure, which is typical for many
mucins (Fig. 5). The predicted molecular mass for the protein from the coding
region is 42,950Da, indicating extensive posttranslational modifications to produce
the mature protein of 150kDa. RT-PCR confirmed that 171.-7 is expressed in
salivary glands and hemocytes, just as PNA staining suggested for p150 (Fig. 5 B).
The lack of signal in other tissues shows thatITI-7 is not ubiquitously expressed. A
second glue protein (SGS-5) was only expressed in the salivary glands, a pattern
expected for a typical glue protein. It can be concluded that p150 is 17l-7 , and that
this protein is expressed in both hemocytes and salivary Qabial) glands. This pattern
of expression, while surprising, is not unique, as hemomucin is also expressed in
both Drosophilahemocytes and salivary glands (Theopold et aL.,1996; Theopold e/
aL.,2001).
To test whether immune proteins have been co-opted by the labial glands in other
insects, the expression of one of the major products was examined (the fibroin
heavy chain, F-hc) of labial (silk) glands in the two lepidopterans Galleria and
Ephestia. The Galteria ftbroin heavy chain sequence was known (Zurovec and
48
Sehnal, 2002), while in this laboratory several ESTs coding for the Ephestia fibrorn
heavy chain had been previously isolated (see material and methods).
RT-PCR results show that both Galleria and Ephestia F-hc are expressed in
immune tissues in addition to their expression in silk glands (Fig. 6 A and B)'
While Ephestia F-hc is expressed in hemocytes (Fig. 6 C) similat to I7l-l and
hemomucin in Drosophila, Galleria F-hc was not expressed in hemocytes, but in
the fat body (Fig. 6 D), which is also an immune tissue.
Discussion
This chapter presents evidence for an immune function for some labial gland
proteins. Previous observations suggested a dual expression of several proteins in
both hemocytes and labial glands. One of these proteins is hemomucin, a cell
surface mucin, which was first isolated from a Drosophila hemocyte cell-line and
later found to be strongly expressed in salivary glands (Theopold et aL.,2001). In a
screen for hemocyte proteins, several ESTs coding for fibroin heavy chain had
previously been isolated and shown to be one of the major products in the silk
glands from Ephesria (unpublished data).
Here it is shown show dual expression in labial glands and hemocytes for a
secretory Drosophila mucin, which is released from hemocytes upon bleeding and
involved in subsequent clot formation. The novel mucin turned out to be identical
to a previously described salivary gland protein (l7l-7). The fact that glue proteins
are also involved in immune reactions could constitute an interesting paradigm.
Either the salivary gland proteins where co-opted by the immune system or,
alternatively, the glue proteins evolved from immune proteins. Although it is
difficult to decided at this stage, which of the two scenarios is correct, the fact that
49
the immune system precedes salivary gland functions, it is likely that the putative
immune function of l7l-7 is ancestral to the labial gland function. This led to
revisit previous work on silk glands and confirm at a molecular level that fibroin
heavy chain is expressed in both the labial glands and in immune tissues in
Ephestia and Galleria.
In both Drosophila and Galleria I was able to observe entrapment of bacteria in the
clot, which could be stained in both insects using PNA. The specific involvement of
171-7 was shown using bacterial binding assays (Fig. a). In these assays, the
interaction between bacteria and 17l-7 was inhibited in the absenc e of C** and in
the presence of sugars, which mimic the determinant recognized by PNA. Two
other mucins present in hemolynph did not bind to bacteria. These data indicate
that both carbohydrate determinants on l7I-7 and a (most likely humoral) C-type
lectin activity may participate in crosslinking bacteria to lll-7 . Amongst the C-type
lectins identified in the Drosophila genome, there are in fact several interesting
candidates for clotting factors including hemolectin (Goto et al., 2001), which
shows domains with similarity to von Willebrand factor and a humoral lectin,
which is in fact specific for galactose (Haq et al., 1996). In lepidopteran
hemolymph plasma PNA-like lectin-isomers are the most abundant lectins.
Attaching bacteria to the clot prevents their dissemination into the circulation and
may be important for immune defense.
While I was able to confirm that both Ephestia and Galleria fibroin proteins are
expressed in immune tissues, differences should be noted between these two
species. In Ephestia fibroin is expressed in labial glands and hemocytes (Gillespie
et al., 1997),l1ke l7l-7 in Drosophila, whlle in Galleria, fibroin is expressed by the
fat body instead (Elrod-Erickson e/ aL, 2000). While both of these are fibroin
50
proteins, they show limited homology (only Ephestia fibroin contains short
stretches of collagen-like repeats). Molecular differences are even more apparent
when both fibroins are compared with I7l-7. These differences in structure and
expression might partly account for the morphological differences Gregoire
described for clots (Gregoire, 1974) and are likely to produce matrices with
different physical properties affecting adhesiveness, elasticity and tensile strength.
The biochemical basis for the dual usage of proteins in labial gland secretions and
during hemolymph clotting lies in their tendency to easily precipitate from an
aqueous solution and establish an extracellular assemblage. The same argument
applies for the gut and for the follicle cells in the ovary, where components of
extracellular matrix (the peritrophic membrane and the chorion respectively) have
to harden quickly and where factors implied in immunity and coagulation are also
detected (Theopold et aL.,1996).
In conclusion, these hndings indicate that labial gland products are used in other
insect tissues, where large assemblages, such as clots, glue or extracellular matrix
has to be established quickly. This is particularly urgent during hemolymph
coagulation both to avoid loss of fluid and prevent infections. In this context, it is
natural to expect that hemolymph coagulation leads to entrapment of bacteria (see
Fig. 3). I believe that this process relies on a number of components, which - like
p150 - are expressed constitutively. In addition, tethering bacteria to the clot is
expected to render them more accessible to other effector molecules of the immune
system, including antimicrobial substances and phenoloxidase, which are activated
at different levels. 'With a better molecular understanding of coagulation, it will be
possible to characterize the specificity of the interaction between microorganisms
51
and clot components and to further our understanding of the interaction between
hemolynrph clotting and other immune responses.
52
Figures
53
(!$ "->\
C}
¿if'.9Q;
" iÌ..1È.'\ir{.
-B ,,1Þ.1',}
ô
'.ì¿(i'
,t ô
Fig. 3: Hemolymph clot formation in two lepidopteran species and Drosophila and entrapment
of bacteria in the clot. Hemolymph from Galleria mellonella (A-C) Ephestia kuehniella (D and
E), and Drosophila melanogaster (F-H) was bled onto a drop containing GFPlabeled gram-
negative bacteria showing entrapment of bacteria in the fibrillar clot. Bacteria are visualized bygreen fluorescence in C, E, and H. -lhe Galleria and Drosophila clots were visualized bylabeling with TRlTC-conjugated PNA in A and F. In I and K, Drosophila hemocytes were bledinto anticoagulant Ringer and permeabllizedto detect intracellular PNA-reactive material.
OCoc oC9L
Sg Hc Br ld Gu FbA
I
; ¡f,
rd
b
250
-98
-64
50
+-qa
B
250 -rso - bl100 -75-
ps ps
üÞl+'J }f
tftl¡*.¡ þl
I
c
-Ca
+Glu
+Lac
+Suc
H
50-100
80
Fig. 4: p150 from hemocytes binds to bacteria. Lysates from different larval tissues were
analyzed in a Western-blot using peroxidase-conjugated PNA (A). The amount loaded
corresponds to 1 animal equivalent (AE) except for the Hc sample (10 AEs), for the silkgland sample (0.1 and 0.05AEs), and for the brain sample (1 and 2 AEs), (Sg=salivary
were incubated with a hemolymph lysate washed extensively to remove unbound material
and the bacterial pellet (p) as well as the supernatant (s) containing unbound material
analyzed on a Western-blot developed with labeled PNA (B-C) or HPL (C, lower parl).
Bacteria alone were included as a control (b). In C, the binding was performed in Ca2*-free
buffer (-Ca) or in the presence of different suga.rs (glucose, lactose and sucrose, each at a
conc. of 200 mMol). The blot from the incubation with sucrose was re-incubated withHPL to detect two additional hemolymph mucins of masses 100 and 80 kDa, which didnot bind to bacteria. Incubation with HPL was performed for all blots were p150 bindingto bacteria had been observed and in no case was any other mucin found binding to
Fig.5: p150 is I7I-7. A: amino acid sequence of l7I-7. The aminoterminalprotein sequence obtained from p150 is underlined. An arg-rich domain and twotypes of triple-repeats are shown. B: RT-PCR detecting expression of Ul-1 an
un¡elated second glue protein (SGS-5) and RP 49 as a loading control in RNAsamples from hemocytes (Hc), salivary glands (Sg) and the brain (Br). For allsamples, 30 PCR cycles \üere run.
D--
c hc fb sg
hcs fbh
Fig. 6: Fibroinheavy chain from Ephestia and Galleria are expressed in immune tissues. A and B:
Phase contrast (A) and in situ hybidization (B) using the insert from an EST coding for the
Ephestia fibroin heavy chain. Strong hybridization was found to the posterior part of the silkgiands (Sg, fat body=F6¡. C and D: expression of the Ephestia and Galleria fibroin heavy chain
was analyzed in the hindgut (hg), silk gland (sg), fat body (fb) and hemocytes (hc) using RT-PCR
with primers for Ephestia ftbroin (a: 35 PCR cycles) and ubiquitin (b: 40 cycles) as well as
Fig. 9: Relationship between members of the SSI family. The same sequences as in Fig. 7and sequences from C.elegans (C.s.), C.roseus, (C.r.) strictosidine synthase as well as
Arabidppsis sequences were analysed for their possible phylogenetic relationship(unbalanced display, where branch distances correspond to sequence divergence) using the
CLUSTAL program as part of Megalign. One of the conserved parts (position 155-179 inFig. 7) is shòwn in (A). Some of the sequences are identical in this part but differ otherwise.
The phylogenetic tree in (B) is based on the sequence between position 116-281in Fig. 7.
GLHGVHGfH
f¡
trF
f'
t,
I
-' r, r'Sir
lË
t' L
/\ D -i\
T] A
LI
l)r'
l,
L
I: J(l l-(ì L
T- (i l
L,lt:trnI] L,
sm sp pa lu li ki br
2.41.4
28S
Fig. 10: Expression pattern of the human SSI member. Total RNA from brain (br),